THE

AMERIOAN JOURNAL OF SOIENCE.

EDITORS

JAMES D. AND EDWARD S. DANA. J I ASSOCIATE EDITORS Pa0FE880BS JOSIAH P. COOKE, GEORGE L. GOODALE AND JOHN TROWBRIDGE, OF CAllBaIDGB.

PKOFE880B8 H. A. NEWTON AND A. E. VERRILL, OF NEW HAVEN,

PBOFE880B GEORGE F. BARKER, OF' PHILADBLPWA. 1 THIRD SBRIBS.

VOL. XXXIX.--,-[WHOLE NUMBER, CXXXIX.]

Nos. 229-234.

JANUARY TO JUNE, 1890.

WITH YUI PLATES.

NEW HAVEN, CONN.: J. D. & E. S. DANA. 1890. Brush and J}ana-Kine1'al Locality at Branchville. 201

urnes, and enough water to increase the total volume to 100 cm', or a little more. A platinum spiral is introduced, a trap made of a straight two-bulb drying-tube cut off short is hun~ with the larger end downward in the neck of the flask, and the liquid is boiled until the level reaches the mark put upon the flask to indicate a volume of 35 cm'. Great care should be taken not to press the concentration beyond this point on ac- o count of the double danger of losing arsenious chloride and setting up reduction of the arseniate by the bromide. On the other hand, though 35 cm' is the ideal volume to be attained, failure to concentrate below 40 cm' introduces no appreciable error. The liquid remaining is cooled and nearly neutralized by sodium hydrate (ammonia is not equally good), neutraliza­ tion is completed by hydrogen potassium carbonate, an excess of 20 cm' of the saturated AolutlOn of the latter is added, and the arsenious oxide in solution is titrated by standard iodine in the presence of starch. With ordinary care the method is rapid, reliable and easily executed, and the error is small. In analyses requiring ex­ treme accuracy all but accidental errors Illay be eliminated from the results by applying the corrections indicated.

ART. XXVI.-On the Mineral Locality at Branchville, Con­ nectiaut: Fifth Paper; by GEORGE J. BRUSH and EDWARD S. DANA. With analyses of several manganesian phos­ phala,. by HORACE L. WELLS.

IT is now nearly twelve years since we published onr first paper* upon the Branchville minerals. It" will be remembered that the material which formed the basis of onr early work was that which Mr. Fillow had brought to light in his excavations, some two years previous, in search for mica. I t was this lot of minerals, sa~ciously selected and preserved by Mr. FilIow, that we found 80 remarkably rich in phosphates of manganese, including a number of new and interesting species.t During "See this Journal, xvi, 33, 114, July and Au~ust, 1878 j Second Paper, xvii, 859,1879; Third Paper, xviii, 45, 1879; Fourth Paper. xx, 257, 1880. t It may be userul here to reCl!1l that our investigations have shown that, besides apatite, amblygonite and some others, the following phosphates are present: Lithiophilile: a manp;anese triphylite, essentially LiMn PO. j maB8ive, cleavable, of a salmon-pink to clove-brown color. Eo8phorite: a man~anese cbildrenite, essentially MnAI(OH)PO.+2H.Oj in or­ thorhombic crystals and massive, of a rose-pink color. 1'riploidite: near triplite, essentially Mn(~{nOH)PO. in fibrous aggregates and monoclinic crystals isomorphous with wagnerile. DidciflJlMlite: a phosphate or manganese, Iron and sodium; in bri~ht green chlonle.like foliated alCgregates, rarely in pseudo-rhombohedral crystals. 202 BrwJh and Dana-Mineral Locality at Branchville.

the yea1'8 of 1878 and 1879, we carried on a somewhat extended search for these minerals in the ledge from which they had been obtained, but the spot from which the most interesting .specimens had been derived was very unfavorably situated for work, being ten feet or more below the level of the ground, and ~ur efforts were only in part successful. Some of the results we have already announced in subsequent papo1'8. Perhaps the most important result of our early explorations was to prov~ the presence of large amounts of potash feldspar' {microcline) and quartz in the vein-in fact, before we ceased ~ur private work, we had brou?,ht to the surface several hun­ dred tons of these mineralB. 'Ihis material was of so excellent (]uality for technical use and the supply seemed to be so large that negotiations were presently entered into between Mr. Fillow, the owner of the property, and the Messrs. Smith, of the Union Porcelain Works, of Greenpoint, New York. with the nnal result of the sale of the property to the latter gentlemen. This was accomplished in 1880. Since that time the work of quarrying for feldspar and quartz has been carried forward uninterruptedly and with gratifying success; up to the present time Mr. Fillow informs IlS that from three to four thousand tons of feldspar and fOllr thousand tons of quartz have been shipped from the locality. The arrangement has proved also a very successful one from a scientific point of view. The Messrs. Smith have very liberally placed at our disposal all the material obtained from the locality which was of no technical value, while the daily presence of Mr. Fillow, with his active interest and keen eye, has resulted in saving for science prac­ tically everything which the locality has yielded. The cov­ ering of earth was early removed, and the ledge opened to as ~l'eat a depth as the drainage would allow; since then the drain has been repeatedly cut deeper until ill the summer of 1888, ten yell1'8 after onr fil"8t work, the time to which we had been constantly looking forward arrived and the deep spot from which the first supply of phosphates came was reached.* In the meantime, however, the work had not been unpro­ du~tive, and the contents of our third paper upon certain deposits of lithiophilite, eosphorite and other associated min­ erals, and of our fourth paper upon the spodumene and its Reddingile: (Mn,Fe), PO. + ;{ H,O; in pinkish orthorhombic crystals near acoro· .lite in form, also in I(runular ",nsses. Fairfieldite: a phosphate of manganese and calCIUm (Ca,Mn),P.O.+2H.Oj tril'Jinic, usually in folia led masses, of a white or yellowish color and pearly to adamantine luster. F'iUowlte: a phosphale of manganese, iron, cnleium and sodlumj in granular aggregatl's of monodinic crystals of a honey-yellow color and resinous to ada· mantille luster. • It may be added that at present the depth of the opening is some 40 feet and is length and breadlh about 160 Ly 45 feet. Brush and IJanar-Kineral Locality at Branchville. 203 alteration.products, show in part what was accomplished. In addition to what is mentioned in these papers, the locality has at several different times yielded a not inconsiderable amount of uraninite, in part in octahedral crystals with a specific gravity of 9'3; this has been investigated chemically by Com­ stock.* With the uraninite have been found two or more­ uranium phosphates which have not as Jet been thoroughly studied. Columbite has also been found in considerable quan­ tity, aggregating more than 500 pound~. This occurs in crystalline masses, and in part well developed crystals and ~oups of crystals in parallel position of remarkable size. It has a specific ~ravity of 5'73, and as shown hy an analysis by T. B. Osbornet contains 19'~ per cent of Ta,O.. Another kind of columbite has also been found in minute reddish brown translucent crystals usually implanted upon the spodumene.:j: This variety Comstock lll\s shown to be exceptionally inter­ esting in the fact that it contains manganese with practically no iron, and further has the niobium and tantalum in the ratio ?f 1: 1; it has a specific gravity of 6'59. Other pointa of mterest that have been brought out are the occurrence on a rather abundant acale of a mineral, both massive and indistinctly fully crystallized, which resembles cyrtolite but has not yet been investigated; also of smoky quartz, in part well crystallized, and remarkable for its riehne&1 in fluid incll1sions (CO., etc.) as described microscopically and chemically by Hawes and Wright;§ also of beryl in large columnar masscs sometimes two feet or more in len~th; still further of albite in finely crystallized specimens. Apatite has been found in a variety of forms; one variety, of a dark bluish green, has been found by Penfield II to contain 10'6 per cent of MnO. Other kinds are interesting crystallographically and resemble the Swiss crystals in habit and complexity. Mica has been obtained in limited amount of a merchantable form (300 pounds of plates cut to pattern); the most common variety, however, is that occurring in curved plates, presenting a smooth convex surface like a watl~h-glasa; these aggregates have a radiated as well as concentric structure. Specimens of the Branchville mica have­ been analyzed by Rammelsberg.,. The most important developments, how~ver, have been those of the summers of 1888 and 1~89, when considerable quautities of the manganesian phosphates were brou~ht to light. This result has beeu especially gratifyin~ to us, smce it has ~iven us specimens of all but one of the new species described III 1878 (cf. p. 202), several of which we had almost despaired of • This Journal, xix, 220, 1880. t This Journal, xxx, 336, 1886. Ihid., xix, 131, IS80. ~ Ibid.• xxi. 203. 209, 1881. fIbid., xix, 361, 1880. ~ Jahrb. Min., ii, 224, 1885. 204 Brush and .Dana-Mineral Locality at Bl'a/lchville. finding again. It has also afforded another new member of the triphylite gro.up, a sodium-manganese phosphate, which we shall call natrophilite. Besidcs this \ve have identified another phosphate of manganese, and one which from the first we had hoped to find, viz: the rare mineral hureaulit:e, thus far only certainly known from Limoges, Commune of Bureaux, in France. The geueral method of occurrence of the phosphates of manganese is snch as to confirm the opinion that we have expressed in a former paper, that the manganesian triphylite or lithiophilite is the parent species. This is beyond all doubt an original mineral in the vein, occurring intimately associated with the albite, quartz and spodumelle. With it, sometimes entirely enclosed by it, we find another of the Branchville species, triploidite, which seems to be also an original mineral. The masses of lithiophilite are as a rule unaltered, showing only a little oxidation on the outer surfaces, and on the fracture surfaces an occasional thin film of a bright blue crystallized mineral in minute amount, resembling dvianite. In some few cases they llave a distinct though rough crystalline form, and the planes 110, 120 and 021 have been identitied; one of these rough crystals is no less than eighteen inches in length and weighs fifteen pound!l. The angles measured are mere approx­ imations, but they are interesting as showing, as has been assumed, that the form of lithiophilite is the same as that of triphylite. We may add here that a 6mall crystal fonnd some 'years ago resembles fig. 450 in Dana's System. L1tblopblllte. Trlpbyllte (T8cbermlk). 110",110= 48° 41° 110", 120 = 17 _18° 11 31' III 0 ~ 021 = .j2 43 31 The cleavage angle 110", 1I0 was found to be 48° 30' to 49°. Occasionally the masses of lithiophilite are extensively chan~ed; this is true especially of some of thc large nodular inclUSIOns in the vein. It is most interesting in these masses to note the change from the perfectly fresh lithiophilite through a zone in which the green chlorite·like dickinsonite is more or less dis­ tinctly represented to the hureaulite beyond. This change of the lithiophilite into hllreaulite can often be obsen'ed, and the interpenetration of the two minerals is well shown under the microscope in sections cut parallel to the basal cleavage. A thin line of eosphorite often crosses the hUl'ealllite. Some­ times with the dickinsonite we have also fairfieldite in clear foliated mllsses, and reddingite ill light pinkish crystals and crystalline masses. It is impossible to lay down any general rules in regard to the order of occurrence of these minerals or their method of arrangement. On the contrary they occur

yGooglc BrU8h and Dana-Mineral Loeality at Branchville. 205 eloeely together, now one and now another prominent, and it is not uncommon to find five or six of the phosphates in char­ acteristic form ou a single hand specimen a few square inches in surface. In one sncb specimen, for example, we have at -one extremitJ lithiophilite and triploidite, then hureaulite and natrophilite interpenetrated by a narrow zone of eosphorite and again lithiophilite. In another specimen which gives a section into the interior of a nodular mass, we pass from the -exterior albite through a nondescript zone blackened by oxida­ tion and showing some dickiJlsonite, through distinct foliated masses of the same mineral enclosing patches of natrophilite, then hnreanlite abundant with fairfieldlte and some reddingite, and finally the unchanged lithiophilite. Much of the material is of the most ill·defined and problem· atical character, being an intimate mixture sometimes of eos­ phorite and quartz, such as we have before described, or of others of the phoaphates in indistinct form. Typical specimens of the different species are as characteristic as could he desired and recognized at once; in others less characteristic they resemble one another so closely that identification is often a matter of much difficulty. It Illay be added that much of the triploidite can hardly be told by the e,y,e from a massive red ~met commonly associated with it, whIle the fairtieldite sim­ ilarly resembles the albite; in these cases, however. a simple test of hardness is all that is required. We go on now to a detailed account of some of the more interesting species identified. N ATROPHILlTE. The sodium-manganese member of the triphylite group, to which we give the name natrophitite, has been identified only in the material obtained during the la!lt summer. It occurs sparingly, usually closely associated with lithiophilite, and upon a superficial examination could be confounded with it, although {}istinguishing characters are not wanting. It appears in cleav­ able masses for the most part, the cleavage surfaces often broad, and showing something of a pearly luster. Occasionally smaller grains appear imbedded in the cleavage Iilass, and these show at times a more or less distinct crystalline form. On one of these the usual planes of triphylite were identified, 110, 120, ~21, 001 (cleavage). The angles could not be obtained accu­ rately but were sufficient to determine the forms, viz: Natrophlllte. Trlphyllte. 110" 110 = 60° 30 47° 1:10" 120 = 87 82 l' 001 ... 032 = 47 -49° 4629 In crystalline form, then, it agrees as was to be expected with triphylite and lithiophilite. Optically it also corresponds so far 206 BrUlJh and lJana-.31i/lei'al Locality at BrancnviUe.

as it has been investigated; the optic a..-xes lie in basal section and the acute bisectrix (po!'itive) is normal to the brachy­ pinacoid. The characteristic basal cleavage is always a promi­ nent character, but the brachydia~onal cleava~e 010 is le88 distinct than is shown by lithiophilite, and the prismatic cleav­ age (110) is interrupted; tIle measured angle was 50°; these efeavages are seen more clearly in thin section". The fracture is conchoidal, more perfectly so than with lithiophilite. The color is a rather deep wine-yellow, much like that of the Bra­ zilian topaz. The luster is brilliant resinous to nearly adaman­ tine; it was, in fact, the brilliancy of the luster which first attracted our attentiou, and which is, so far as the eye is COD­ cerned, its most distinguishing character. The mineI'al itself is perfectly clear and transparent, but the masses are much fractured and rifted. The surfaces are ofteu covered by a very thin scale of an undetermined mineral, haviug a tine fibrous form, a delicate yellowish color and silky luster. This same mineral penetrates the masses whereYeI' there is a fracture surface of cleavage or otherwise. What the exact nature of this mineral is we are unable to say, since the amount is too small to admit of a satisfactory determination-it appears to be a man~anesian phosphate. It is evidently an alteration­ product and would seem to imply that natrophilite is rather subject to easy chemical change. In any case this silky film is one of the characteristic features of the mineral, and directs attention to it at once even over the surface of a hand specimen where it is associated with Iithiophilite and perhaps three or four other of these phosphates. Before the blowpipe natrophilite fuses very easily and colors the flame intensely yellow, thus being at once distinguished from lithiophilite. It also gives the usual reactions for manga­ nese. The following is an analysis of natrophilite made by Wells. The specific gravity on two fragments was found to be 3'40 and 3'42,

I. II, III. Mean. p.O. __ 41'03 41'0:1+142= '289=1 =1 MnO .• 38'19 38'19+ 71 ='5:181 '580=2'01=2 FaO __ . 3'06 3'06+ 72='042 f Na.n._ 16'77 16'81 16'79+ 62='2711'279=097=1 Li,O _. 0'20 0'19 0'19+ 23='008 f H,O. __ 0'40 0'45 0'4:1 Inaol. _ 0'81 0'81 0'81 0'81

100'50

I II I n The formula is therefore H.O. 2RO, P,O. (lr RRPO., or essen- tially NaMnPO.. It will be noticed that iron is present in very small amount only (3 p, c.) and of lithia there is hardll more than a trace (0'2 p. c.), With the discovery of natrophi - B1'U8h and Dana-Mineral Localdy at BranahviUe. 201 ite, the triphylite group receives an important addition, and we now have: Triphylite, LiFePO. ~ Connected by mSllY intermediate Lithiophilir.e, LiMnPO. f compounds, Li(Fc, Mn)PO •. Natrophilite, NaMnPO•. These three species are, as is to be expected, closely isomor­ phous. To them is also related in composition and in some degree in form the new sodium-beryllium phosphate, bel'yllo­ nite, NaBePO., which was described"by one of us a year and a half ago.* The relation of natrophilite in origin to the common lithio­ philite is an interesting question. In view of the extensive changet! that, as we have shown, have taken place in the spo­ dumene, by which the lithium has been removed and its place taken more or le88 fully by sodium, or sodium and potailSium, it is natural to suggest that a similar change has resulted in forming the NaMnPO. oot of LiMnPO., and this we regard as very probllble. Its limited method of occurrence suggests the same thing, although it must be remarked at the same time that it seems to pailS into hureaulite as readily as the lithiophilite. If in fact formed from lithiophilite, the change probably took place before the formation of most of the other phosphates.

HUREAULITK. Perhap's the most interesting of r~cent developments at BranchVIlle is the discovery of the rare mineral hureaulite. Thus far our knowledge of hureaulite has been limited to the account of crystals from Limoges by Dufrenoyt and the later and more thorough description by Damour and DesOloizeaux·t In addition we have only the single remark by Websky that it probably occurs at Michelsdolf, Silesia, with sat·copside. The crystals described by DesOloizeaux belong to three varieties showing two distinct types of form. though having the same composition, as shown by Damour. These varieties are respectively violet-rose, brownish orange and pale rose pink in color. Their crystallographic relation to each other is anoma· lous, in fact, it would be difficult to find another case equally so. The crystals of the two types have the fundamental prism in common, bot otherwise no plane of the one occurs on the other, and what is more remarkable, the symbols assigned to a number of the planes of the second type are complex in the extreme. The axial ratio calculated from DesCloizeaux's fun­ damental measurements is a: b: c= 1'6977: 1: 0'8887 j f3 = 89· 27'. The planes observed on crystals of the two types are as follows: • Thia Journal, xxxvi, 290, Oct., 1888; xxxvii, 2'1. Jan, 1889. t Ann. Chim. Phys., xli, 338, 1829. *Ibid, III, liii, 293, 1868. 208 B'I'U8h and Dana-Mineral Locality at Branchville.

~--Flnttype --~ ~---Second type"--"::"n-••"='ci. n..CI. b 010 i·i g a 100 i·i hI c 001 o p m 110 1 m m UO I m o 106 -!·i o· g 301 -3-i a 15 . 0 . 8 ~f·i Q1"- 011 l·i d 435 -H d u 311 3·3 k 19' 6 . 8 '.B·ll k 341 4.: x II . 9 . 10 H·.Igl x £ 9. 11 • 10 H-,y. ~ The Branchville crystals, like those from Limoges, vary in -color from pale violet to reddish brown and deep orange-red. The habit of the crystals, however, is nearly constant and the angles also so far as our measnrements have gone; they corres­ pond to the second type of the Limoges crystals. The cryBtalH are not easy to decipher, since they are vel'y small, united by parallel grouping and as a rule present only a few planes in ~mch a way as not to exhibit the symmetry. The angles are not as accurate as could be desired, although the crystals are much better than those of Limoges, since DesCloizeaux givee his observed angles to whole degrees in many cases and the majority are stated to be approximations only. For the sake of greater simplicity of symbols, the position of DesCloizeaux is modified somewhat in that his plane 105 (0·) is taken as the base and the pyramid ais made the unit pyramid. For fundamental angles the following have been assumed: 100" 001=84° l' 110" 110=62' 21' 110,,401=70' 64' whence we obtain: a:b:c=1'9192: 1: 0'5245; 13=84' 1'. The observed planes with the symbols of the corresponding planes so far as observed by DesCIoizeaux are as follows: ne.ClotJeaox. a 100 i-i 100 hI cOOl 0 105 05 m 110 I 110 m a ~01 4·i 15'0-8 a,66 13 50 I 5-i P 223 -t d 111 -1 435 e 221 2 9'11'10 e k fill 5·il i95'8 k z 621 6-3 I g,B 8-2 The attempt to transform the symbols of DesCloizeanx, according to the usual methods, into those required by this -change of position meets with only partial success. Thns the plane 19'5'8 becomes by the transformation 411 while the observcd a~les of DcsOloizeaux make it for the axial ratio here taken 511. As will be seen below the anKles of anum' ber of forms on the Branchville hureaulite agree pretty well with the angles measured by DesOloizeaux, with the single exception of the prism. For this he found 119°, while we make it for the Branchville crystals 124° 42'. It is this dis· crepancy which caUBeS tIle want of agreement which we have just alluded to. !<'urthermore, it is seen that the complex symbols of several of the planes, in DcsOloizeaux's position and referred to his axes, become simplified when referred to the axcs here adopted. Of the .I?lanes noted by DcsOJoizcaux on type 2, all but one (II'9'1O) It will be scen, occurs on the Branchville crystals and to this the symbol l)82 probably be· longs in our position. Of the forms of the first type, only the prism occurs with us, but to the other planes thc probable ~ymbols in our position may be assigned. DeoC'ol••alllf. V..Clolr.aus. 0(11 11 lOB all u 12'3'27 OJ! c' Ill:! 9.1 I i10IT TIle following table gives the lUore important anKles calcn- _I,.,Clol.f&us._-. ~ifci:--uraDch'll~.~lIrod. CalcUltlld. )I...ur.d. {02" 111',02' :W, ~ 100 AlIO 002' 21' 02 27 02- 811 liD' 110° 02 411 ~01 A 100 os. I 83 20 83 211' 83' 80' u rOO 911 119 9el 9 90 81 90 80 .. .03 II tal (II 118) .. -lOI 110 49 110 .2 110 110 100", .}Ol 411 10 4Ii II~ 411 49 4Ii 40 100 A 228 74 47i .. III 71 211 71 9 Ioo .. kl 47 119 .. A&11 41 .0 42 12,42 14, t42 811 ~ n 10 001 A228 21 3 21 8,21 118 .. Alii 20 40 20 40,29 28 30 40 20 211 .. A 110 87 14 t87', 87· 4', 80 42 80 87 12' ~ .. 221 61 17 110 17 110 29 u A&11 69 80 119 80 .. ,,110 92 .0 92' 110' 2211 A 223 87 9 37 87 111 '" I 1 112 14J III 114 112 IS III 1111 A&11 811 118 80 7,80 81 88 88 1n 110,,111 117 28 117 82,117 22, 110 2 1111 41i t67 .2 .. 223 00' 11' Oil 12,07 8 f 110 221 41 29 42 I 48 2 48 liD, .. ,,&11 III' 9' 40.111 48 117 t'll61 62 .0 .. AA41 26'13' 211 13, 211' 22' f .. ,,401 °70'114' 70 113, H' 19'

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. lated from our axial ratio compared with our meallurements and also with the 1Tl6lU'U1'ed angles of DesCloizeaux. It, i~ seen that the correspondence between our measured anll calcu· lated angles is not in all call88 l\8 great all could be desired, although as much as could perhaps 00 expected from the nature of the material. It has been stated that the crystal8 are often grouped in parallel position, bnt all is common in such caaea, the parallolism is not perfect and furthermore the part8 8how sl~ht variations in position, even when tho planes arc smooth, which are doubtless to be referred to the Bailie C8UBe. The habit of the Branchville crystals is short prismatic 88 .hown on f1llure J: a basal Droiection of a more comnlex form ill gh'en In fig. 2. The groupin~ in parallel p 0 s 1 tion gives rise to a repe· tition of tho pria­ matic plane8 which ma, l'OIJult in a deep .. 8trlation or furrow· ing of tllis face. Besides this, the %one of Dlanes m. 1, k, a, is often striated or channeled pa1'll1lel to their common direction of intersection. The crystals show rather perfect cleavage parallel to tho orthopinacoid, For analysis, carefully selecte{l portions of the purest crystals were taken by Prof. Wells. Tho sJ>CCific gravity was found to 1>t' 8'149. The results are 8l\tisfaetorJ' lIB aKrecing fu!!,y with those of Damour and leading to the same formula. For comparison we quote Damour's analJ'sis of the palo roso crystals which ditlers but, little from that of the yellow crY8tals; it i8 to be noted that the violet crystals (tl.p8 I) ha\'e not bee" analyzed and it is pouihlo that somo dltlorenco in cOlllllosition may explain the ditlerenee noted in the, forlD. The Branchville mineral con­ tains a little lese iron than the Limoges.

BranchTllle, 0,-..3'149 L JI, M_ PlOt ••••••• ~8'28 1l8'.. 38'38 'S10=1'OO=2 37'83 FeO •••••••• 4'1S 4'31 4'58 '083} 8'18 MoO ••••••• 42'29 42'29 '598 '818=2'50=5 41'8D c.O ••.••..• 0'94 O"N 0'11 R.O .••••••• 12'U 12'15 12'20 '67'=2'01=4 H'60 Qua"••••••• 1'78 1'1S O.Dp' 0'30 100'11 100'2&

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The formula is lSRO. 2P.O•. lSH.O or H.R.(PO.).+4H.O. The percen~ composition calculated for manganese only is: p.O. 88-96, !InO 48'69, R.O 12·8lS=100.

RBDDINGITB. Tho species reddin~ite !las been known thus far only in a few specimens, sbowmg it in a granular fonn of a reddish ~olor, or rarely in octahedral crystals often superficially black from oxidation. The material first found though scanty WliH .ufficient to admit of tIle determination of t)lO form, wlJiclJ was ,shown to be similar to tbat of scorodite and 8tren~ite. Among tbe specimens recently discovered, reddingite IS not uncolD­ man, and we have been gratified to obtain it not only well ~rystal1ized but also in mfUlBive form, perfectly fresh and unaltered. The color is a pale rose-\linlC, often hardl~ moro than a pinkish-white. The most intimately associatea min­ erals are fairfleldite and dickinsonite, the latter of which is often imbedded in it in isolated scales or more often in steJlate groups of green folia. The octahedral habit of the crystals, which appear in occasional cavities, is usuaJlv apparent at 1\ glanco, out not infrequently the crystals are ~j8torted by tho ~longation of a pair of pyramidal planes which ~ves them a misleading oblique prismatic appearance. The common form D. ..

of the crystals is shown in fig. 8 from our former paper. Some of the orystals are more complex (fig. 4) and show also the unit pyramids r, 11 and t, whose symbols arc respectively 888, 223, 114. Theile planes do not give sharp measurelllente but the angll'lI are sufficient for identification. The axes taken arc those previously obtained by us, viz:

eI:~: ~ - O'S6~S: 1: O'O(SII and the measQred and calculated angles are, M.....rod, Calntat.d. III .. 311S .. 20· 11' 21t· 0' III " 223 0:: 10· 10' II· 114' 111 " 11' - 10· 41' Ill· 46' It seemed especially desirable to have a new analysis of this species, both because the material was more abundant. and

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. better than wbat we had had before, and also since the compo­ sition-though in fact fully established-may have appeared to some nnomaloulI, in view of its failure to correspond with that of ecorodite and strengite in the de~ of oxidation of the manganese and in the amount of water. The new analysis b1. Wells fully confirms the former one made by him, onlt dltfering in the larger percentage of ferrous iron present. This analysis of a carefully selected portion with a specific gravity of 8'204, gave: I, u. P.O. SC'90 ... 142 = '2C8 ... 1'00 = 1 F.O n-Is ... '12 ....211S} UIlO SC'1l1 ... 'II.'CS6 'Ull =2'99 =3 Cao 0-611 ... 48 ... '011 H.O IS'18 13'lS ... 18. ''IS2 = 2'9S _ a Quarts O'IS 100'C8 The formula is hence R.(PO.).+3H,O, and if R=Fe:Mn... 1:2, this requim p.o. 84'64, Fe017'M, Mn084'68, R,O 18'17=100. F"'laPISLIllTS. Fairfieldite appeal'8 among the specimena recently obtained not infrequently, and in a form much fresher and purer than that in which we bad it before. It is uaually ID foliated masees intimately lUlIIociated with reddingite an(J fairfieldite, and hardly leM so with hureaulite. The color varies from white to yellowish or, Kreeniah white; it is usuaUy perfectly transparent and the luster is very brilliant, varying from adamantino to pearly according to the surface on whicb it is viewed, die latter on the surface of perfect cleavaKe. A tendency to crystallization is at times apparent but no cryatals suitable for measurement have been found, which is to be re~tted since the early results left much to be desired. An analysis of the perfectly fresh mineral haa been made by Wells. This agrees with those of Penfield previously published; the amount of iron is le88 and that of tIle manganese greater but it is wonhy of Dote that dIe ratio of 2: 1 for Oa :-Mn+Fe is still rnaintained.t The analysis of pure material baving a specific gravity of 3'07 is as follows:

• The ..rller analYl1I gin FeO 6'43, 1111046'29, t It II lat6,,"lIog to ('1111 allflltloa here to the Idelltlllcalloll of falrlleldlte by 8alldbt~r at Raben'lflll, Jahrb. ~lIn., I, 1811. IS811, It II ,110 "ortllY of Dote that II Ilew hydroua phOlphate of ferrous lroll aad calcium, Ilear falrfleldlte but with 2".,0, hll be9n rectntlyllllmed m_llte byllulbmaoQ (ZI. Kryat., nil, as. 1889); Ilk., '.Irfteldlte It II trlcllnlo. Furthermore the bralldUte of Nord.ll­ ul6ld II Ca.WIlIA,O.). +28,0, comapolldlar ezact1y to falrfl.ldlte, (Eh'. Ak. Btockh.. 489, 18S8, Groth, T,b. Ueb• .'410" p. SO, 1889.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Brtf,lJh and llana-Mineral Loaal'ity at Branohville. 218

P,O. [31'69*] + 142 = '266 = 1'00 = 1 FeO 3'42 + 72 = '047 t MnO 17040 + 71 = '245 f '292 = 1'10 = 1 Cao 30-02 + 56 = '636 = 2'02 = 2 H,O 981 + 18 = '345 = 2'06 = 2 Quartz 1'66

100,00

The formula is hence essentially Ca 2 Mn(PO.).+2H.O, which requires P .° 5 39'34, Mn019'67, CaO 31'02, H 09-97=100, This analysis confirms the earlier one bl Penfield and further makes.it probable that there is a definite ratio of 1: 2 for Mn (WIth Fe) : Ca. " -

DICKINSONITE, One of the most remarkable and novel of the species first dellCribed from Branchville was the chlorite-like dickinsonite­ a mineral of a bright green color, micaceous strncture and pseudo-rhombohedral form. Recent developments have en­ abled us to add materially to our knowledB'e of the speciel:'. The number of specimens obtained is relatIvely large and in some of them it shows itself in tolerably well-defined crystal­ lized forms. It will be remembered that for our earlier work we had only one or two minute crystals. The habit of most of the crystals now found differs from that before described, fig. 5, the hexagonal form is rather rare and the crystals appear as rectangular tables united in sli~htly diver~ing groups, fig. 6. 6, 6.

A closer examination shows that they agree with the same fundamental form before accepted. These crystals are elon­ gated parallel to the orthodia~onal axis, and the basal surfaces are bent and striated in this dIrection. In addition they show on the edges, sometimes in traces ouly, the pyramidal planes, which when developed give the hexagonal habit before noted. In addition to the planes a, b, c, r.tJ (301),p (111) and 8 (221), we have identified also a steep clinodome, n, which has the symbol (051) and a hemi-orthodome, y (103). We retain our former axial ratio ~: b: ~=1'73205: 1: 1"19S06; p=61° 30'. * By dilference. 214 B1'U8h and Dana-Kineral Locality at BranchviJ,k.

And the measured and calculated an~le8 are: Calculated. Me&llurad, 001" 051= 79° 16' 79" 001" 103= 12 60 13-14 Optically we find the cr,Ystals as before stated to be biaxial, the optic axes baing situated in the clinodiagonal section and the bisectrix nearly normal to the cleavage-face; the double refrac­ t.ion is negative and the axial angle large. Besides the crystals occasionally appearing in the cavities, and often united in slightly diver~ing groups with edges par­ allel to b projecting, the mineral occurs foliated to almost massive and granular, the folia, however, usually distinct and often grouped in rosettes or stellate forms. Dickinsonite is the species about whose composition we felt most doubt wheu we first published. The material then in hand was very scanty and not entirely pure, and although ex­ cellent analyses were made by Penfield, their interpretation was a matter of some doubt because of admixture of more or

DIOKINSONITE, BRANCHVlLLE. Analysis of first sample. Bp. Gr. S-i4S, I. lI. Mean. PIa' 39'67 39'57 + 142 = '279 =1'00==1 FeO );1'25 13'25 + 72 MnO 3\'74 31'42 3\'58 + 7\ = '445 CaO 2'15 2'15 + 56 -= 0'''1'039 MgO trace '814 =2'92=3 NalO 7'47 7'44 7'46 + 62 = '1241 K.O 1'49 1'55 1'52 + 94'2= '017 Lila 0'20 0'14 0')7 + 34 '005 J RIO 1'66 1'65 1'65 + IS = '094 =O'34=i Quartz 2'58 2'58 2'58

99'93 Analysis of s~cor.d sample. I, ll, Mean. P,O, 40'89 40'89 + 142 = '288 =J'OO=1 FeO 12'96 1296 + MnO 31'83 31'83 + 71 =~l"}'448 CaO 2'09 209 + 7266 = '038 MgO none '811=2'82=3 Na.O 7'37 7'3, + 62 = '120 KIa 1'80 1'80 + 94'2= '019 Li,a 0'22 0'22 + 34 = '006 RIO 1'64 1'62 1'63 + 18 = '092 =O'32=i Quartz 0'85 0'79 0'82

99'61 less eosphorite as well as quartz. Two independent sets of new analyses have been made by Professor Wells, The ma­ terial for the first was picked with great care, but in order to B1'U8h and IJana-MineTal Locality at Branchville. 215 remove all question as to whether the results gave the true composition of the mineral, a second and independent analysis was made. For this the ver:! best material was selected and after being separated was minutely examined microscopically to make sure of its purity, The results as will be seen are identical with those of the first. The two samples were picked from separate specimens and the material was apparently very pure. Unusual care was taken in picking the second sample and its purity is indicated by the small amount of quartz present. The formula indicated by both the anal.yses is 3RO. P.O., iR.O or R.(PO.).+iH.o where R=Mn, Fe, Oa, Na., K. and Li,. There is no simple ratio between the alkalies and the remaining bases. The results vary considerably fl'om those of Penfield in his original analysis. This is undoubtedly due to the fact that the present material was much purer than that analyzed by him. Penfield found about 14 per cent OaO, (probably due to admixed fairfieldite) only about 6 per cent of alkalies and 3'87 per cent of R.O. The formula which he arrived at, however, is confirmed except in the amount of H.O. It will be seen that the composition now established is essen­ tially the same with that deduced for fillowite on the. basis of Penfield's original analysis.

FILLOWITE. The fact just stated, that our former formula for tillowite is the same as that now obtained for dickinsonite, has made us very anxious to prove that our early results were trustworthy, especially since the material in hand at the time of our tirst invf'.stigatiou was very scanty. Unfortunately, among the large number of specimens recently obtained from Branchville, we have not succeeded in finding a trace of this mineral. We have been forced consequently to revert to the few originalspecimen8 8till in hand. The best of these we gave to Mr. Wells, and from it he picked out about 0'75 gram, in the homogeneity of which he had entire confidence. .A new analysis of this has been made by him with the following results; for comparison we quote the original analysis by Penfield. Ratio_ Analysts (1878) Penlleld. • p.o. . 39'68 -;- 142 = '279 l' 39-10 FeO __ ._ 9'69 + 72 '135 1 9'33 MuO _ = 39'58 + 71 = '557 f 39-42 eao _ a-63 + 56 = '065 '847 4-08 Sa.O _ 5'44 + 62 = '088 5-74 Li.O ._ 0'01 + 30 = '002 0-06 . R.O _._ 1'58 + 18 = 'Oll8 1-66 Quartz .••• 1'02 0'88

100'69 100'27 Ax. JOUK. ScI.-TmRD SERIES, VOL. XXXIX, No. 231.-MARCU, 1890. 15 '216 A. A. Michelson-Interference Experiment.

It will be seen that the two analyses agrec throughout and the formula, is the same, viz: R.P.O.+tH.O. AP. noted above it is identical with that of dickinsonite, although the latter species contains more alkalies and le88 manganese. 'The two species are then eBBentially dimorphous forms of the same compound, and the relation between them is made all the more interesting in that with the striking differences in physical characters, there is yet an obvious relation in form. Dickinsonite is monoclinic with marked pseudo-rhombohedral symmetry and of fillowite the same is true as we have proved by a reexamination of fragments parallel to the distinct but interrupted basal cleavage. Moreover the dimensions of the forms show a close relation, thus we have:

D1el

ART. XXVII.-A simple lntel:ference Experirnent; by ALBERT A. MICHELSON.

THERE is probably no experiment connected with the wave­ theory of light of greater fundamental importance than the justly celebrated one known as " Fresnel's mirror experiment," and accordingly many find it neceBBary to repeat the experi- . ment for their own edification or for class demonstration. Without the use of rather elaborate and costly apparatus even skilled experimenters find some difficulty in producing the desired effect-the interference of two pencils of light as manifested by the appearance of colored frInges on a screen, or in the focus of an observing lens; and unskilled observers find it almost impoBBible, even with these aids. Even when produced under the most favorable conditions the phenomenon 1S complicated and often much obscured by diffraction effects.